Safety valve noise; limits, reduction and control

By M. D. G. Randall, Foster Wheeler Energy Ltd. Reading, Berkshire, UK.

First a little philosophy

As a contractor's engineer, one wants to have a model or other method of solution in place before one
meets a cause for its use. Surely to have "no available model" shows absence of prior thought. Some
models will show lack of thought. As examples we might think of: a model inconsistent with known
facts or common sense; no data to substantiate the maths; predictions inconsistent with the data. We
will all accept that a simple or basic model is better than no model at all, because, as information is
gathered, the extra descriptions and data can be used to improve or change the model.
In the following discussion the reader will find examples of a simple model, old information, and issues
that are not well defined, but with which it is suggested he work at the present time. No apology is
made for this. The issue of noise from safety valves does not appear to be well covered in the general
literature and by making reference to issues where there is uncertainty, the author hopes that others
may be encouraged to add definition or associate numerical results to them.
What is a safety valve? How often, long and loud is its noise?
The safety valve is a device to avoid a dangerous build-up of pressure within a system that it is
designed to protect. One should never forget that any redesign of the PSV system must not decrease
the safety of the protected system. A safety valve is normally used with compressible fluids, whereas a
relief valve is primarily used with incompressible fluids (see the introduction to Ref 1). The safety valve
is generally known as a PSV and may release the process fluid directly to the atmosphere via a short
stub pipe, or release the process fluid via a pipe to a flare, or some other equipment. These will be
called "open vent" and "closed" PSV systems, respectively. A PSV is actuated by upstream pressure
and is characterised by what is described as a "pop" action upon opening. It is important to recognise
that one should not expect a gentle release of gas proportional to valve lift.
The noise from major PSVs can be expected to be in the region of 150 -170 dB PWL. I will guess a
figure of "once in a hundred years" for the operational frequency of a single PSV, and thus on a plant
with a hundred PSVs a noise from a PSV might be heard once a year.
While the system is depressurising the PSV will make noise. The noise changes and decays with time
as the pressure decreases. The noise is greatest while the pressure drop across the valve induces
sonic velocities in the valve. The higher the pressure ratio the higher the noise. The PSV may "chatter"
due to flow instability while the gas flow continues, and it may not re-seat when the pressure is low
enough for this to happen.
We shall define these emergency releases of gas as transient noise sources, but it may take hours for
a system's total inventory to be released to the atmosphere or to the flare.
What criteria might be used in the evaluation of safety valve noise?
Three noise related criteria are suggested for PSVs,
1.
2.
3.

Noise received at the local community,

Acoustic fatigue of the components and associated pipe-work,
The possibility of hearing damage due to noise received by a worker who is 'close' to an
actuated PSV.

This paper is related to the third of these.

Plant owners' noise limits, for the avoidance of hearing damage risk, from transient sources such as
PSVs, can be expected to be in the region of 100 to 125 dB(A) Sound Level (SL). The 115 dB(A) limit
in API EA 7301 can be considered typical (see pages 16,17 of Ref 2). The sound level is to be
measured or predicted at the worker location (or where he is expected to be), i.e. at ground level, and
on platforms, ladders, and stairs. Sometimes this can be very close to a PSV or its open vent.
PSV noise is radiated from the associated piping and equipment as well as the wall of the PSV. With
open vent systems the majority of the noise exits via the open vent at the end of a stub pipe. In
addition to this, noise may be generated due to the mixing of the high velocity gas with the surrounding
air. PSVs with stub pipes to atmosphere might occur with steam, air, or nitrogen, but rarely with
hydrocarbons.

What regulations and standards apply to PSVs and their noise?

Not all countries and standards organisations require that the safety valve be treated as a source of
sound which has to be limited.
As an example of what might later be seen as an enlightened standards organisation, we can look at
NORSOK and its view that noise from PSVs should not be considered during design. (see Section
6.5.2 of Ref 3) [The noise limits shall not apply to design emergency conditions e.g. near safety --valves, firepumps or outdoor areas during full emergency flaring, etc.]
In the original paper a list of UK legislative instruments was presented so that the reader might review
them and answer the question, " Does this document apply to my system and its PSVs? That list
related to noise and more general safety matters. No such list is presented here. It will be for the
individual designer, wherever he is in the world, and wherever the pressurised system is being
installed:
1.
2.

to find the appropriate codes and standards,

to decide how they are to be complied with.

Here are two documents which include noise limits that seem appropriate for any review of PSV noise.
API Medical Research Report EA 7301. (Ref 2)
This document dates from 1973. It describes the situation under discussion today. It set a limit of 115
dB(A) to steady sound, and 140 dB(peak) to impulsive noise. These were based on the data in the US.
OSHA 1970 Act.
86/188/ECC. (Ref 4)
This directive states that if a maximum value of the unweighted instantaneous sound pressure level is
greater than 200 Pa "suitable and adequate" ear protectors, which can be reasonably expected to
keep the risk to hearing to below the risk arising from exposure to 200 Pa, must be used.
It is on this directive that the UK's Noise at Work regulations are based. (Ref 5)
Remember that as designers or contractors what we have to do is
1.
2.

make a prediction of the expected sound, and

have a series of design options ready for the PSV that approaches or exceeds the limit.

How does one calculate PSV noise at known worker positions?

Standard methods are available for calculation for valve noise heard at a distance. Two methods for
calculation of PSV noise are suggested: the new (1995) IEC standard for control valve noise prediction
(See Ref 6), and Sections 4.3.5 and 5.4.4.3 of API 521 (Ref 7), which appears to be based on the
1950s method described by Franken (See Chapter 24 of Ref 8).
The Franken and API methods have been followed by more than one PSV vendor so a simple
explanation of the noise prediction method is introduced here. Sound pressure level (SPL) at a point
30 m distant from a PSV vent of sound power level (PWL) is calculated by use of the equation,

Sound power level is calculated by use of the equation:

The MC2 term is the kinetic power of the choked flow through the valve, and the symbol h is for the
acoustic efficiency associated with the transformation of some of the kinetic power to sound power. M
is the mass flow rate. C is the speed of sound in the choked gas thus, C 2 = k*R*T/ (MW). Franken
shows that h ranges between about 10-5 to 10-2.

With the above two equations the Franken discussion leads to an equation for the SPL at 30 m:
SPL30 = 10*Log[MC2* ] + 79 (3)
Franken provides a graph of h vs. Pressure Ratio.
The API formulation for SPL at 30 m is similar:
SPL30 = 10*Log[MC2 ] + (10*Log[ ] + 79 ) (4)
where the value in dB of the term (10*Log[ ] + 79 ) is to be evaluated from the graph (dB vs. Pressure
Ratio) published in the Recommended Practice.
The ordinate scale on the API graph is labelled 'L = L30 -10Log(MC2)', however the simple derivation
given above displays its physical basis.
The curve that is the result of plotting jet acoustic efficiency, against pressure ratio is shown in Fig. 1.
Both the API, (10*Log[ ] + 79) and Franken, ( ) ordinate scales are given for this characteristically
shaped curve.

Fig. 1. Acoustic efficiency of a choked jet.

At least one PSV vendor has simplified things yet further. By ignoring minor PSVs with pressure ratios
of less than three, and with the assumption that one value of acoustic efficiency, say 4* 10 -3, can be
used for pressure ratios above three, a formula results for the sound pressure level at 30 m which is:
SPL30 = 10*Log[MC2 ] + 55 (5)
Other methods that are used for calculating the sound power level of PSV's can be found in the
technical literature. We had one project at Foster Wheeler Energy where there was a proliferation of
PWL prediction methods and we found it expedient to initiate a labelling system. In this system the
initials of a methods author/s were written after the symbol PWL to indicate the source of the
calculation. Note that no one method appears to have been adopted by the "valve industry" and that
no international standard exits solely for the calculation of PSV noise.
The reader may wish to calculate an example, so a relevant data set is given here,
Mass flow rate
Steam
Ratio of specific heats

M
40 kg s-1
MW 18
k
1.26

Universal gas constant R

8.3* 103 J K-1 (kgMole)-1
Temperature
T
300 +273 K
Pressure ratio
PR 61 : 1
A prediction method for noise at worker positions due to PSV vents was required, and in the absence
of an acknowledged and accepted method, the following simple one was developed to find the sound
level at an angle and distance from a pipe vent. Consideration is given here to the case of an "open
vent" rather than a "closed" system as it will be seen to represent the most onerous case.
It is reasonable to suppose that the noise at the worker positions close to an open vent PSV will be a
function of:
valve body and pipe radiation (themselves functions of frequency)
radiation from vents which are directly to atmosphere ( a function of frequency)
the distance to the nearest " worker position"
any directivity associated with the ratio of, wave length of sound to vent diameter ( the gas may be cold
or hot and possibly about mach 0.5).
Assumptions were;
1. The API 521 calculation provided a dB(Lin) value that could be approximated to the same value in
dB(A)
2. Directivity factors for the stub pipe's vent could be found and used.
3. A peak frequency of noise could be selected so as to act as a guide to the directivity factor.
4. Reduction of noise with increase of distance would follow the inverse square law.
A simplified set of directivity curves was used for the initial trial calculations with this method. They
provided the increase or decrease of sound with angle and were based on data in ISVR course notes (
See Chapter 9 of Ref 9) which was itself based on BBN data from 1952 and VDI 3733 data of 1983.
This directivity data is not given in VDI 3733 of July 1996. This data was reduced to a directivity table
after the selection of a peak frequency to act as a guide to the directivity factor. See Fig 2 for an
example based on 1000 Hz.

Fig. Figure 2. Assumed

directivity (in dB) at 1000 Hz.
Initially, there was a lot a work involved in gathering the data on the positions of the platforms ladders
and stairs, and the positions of the closest PSV vents, but now, with the help of Foster Wheeler /
Intergraph 3D PDS it is a much quicker process.
An example of a PSV noise spreadsheet is given in Fig 3. This spreadsheet shows the "safe distance"
at which a worker might stand. "Safe distance" is defined as the distance at which the sound from the
PSV has reduced to the safe level. The safe level will be the level set by the owner, but is typically 100
to 125 dB(A) SL as was noted in Section 3. The safe distances from the open vent and from the pipe
are different because the rules for noise attenuation with distance from these types of sources are

different. We assume that the sound from the vent falls off at a rate inversely proportional to the square
of distance from the vent. We assume that the sound from the pipe falls off at a rate inversely
proportional to the distance from the pipe. At the foot of the spreadsheet there is a warning of the
results obtained from these simple assumptions. The data shown as entered is not "real" but is realistic
and a "safe distance" can be seen for both vent and pipe radiated noise at a range of sound power
levels and sound pressure levels at 1m distance from both the vent and the pipe wall. The table has
been simplified to the bare essentials. No directivity corrections are made in this example and the
effect of pipe wall transmission loss is demonstrated only by a constant difference of 25 dB between
columns 6 and 7.

Fig. Figure 3. PSV design

summary spreadsheet.
"So you know it's noisy, now what?"
As contractors we expect to purchase equipment that complies with relevant laws, codes, standards
etc. We expect vendors of equipment to comply with such stipulations, to have carried out sufficient
research so that they can calculate the expected noise, warn of any foreseen dangers, and reduce the
noise as much as is reasonably practicable.
Recently, to develop the work described above, Foster Wheeler Energy sought the views of the PSV
vendors on the current limits, and methods of noise reduction and control.
A letter of enquiry was sent out to a number of PSV vendors. An edited copy is provided in the
Appendix to this paper and will probably be recognised in some Vendor organisations. The letter
covered:
Philosophy for PSV noise released directly to atmosphere
Limits to PSV noise
Noise reduction ( the possibility of a quiet PSV (<115 dB(A) at 1.5 m) or PSV and silencer element)
Noise control ( Five possible methods were mooted)
Current method of noise calculation and test basis of noise data
Responsibility for noise prediction
A proposal to provide to PSV vendors design information related to the position of the PSV vents and
manned positions, so that the vendor could use it to advantage in their proposed solutions.
Of the ten companies contacted we received two written replies.
Conclusions
1.
2.
3.
4.
5.

The discharge from a PSV is noisy.

The sound level limits are various, may be impossible to achieve, and pose the question,
How do we demonstrate by calculation that the limits have been complied with?
Some organisations absolve PSVs from compliance with noise restrictions in emergency
conditions.
It would appear that some PSV vendors have yet to provide noise data for their ranges of
safety valves.

6.

Contractors may have understood their obligations but find it difficult to provide accurate
information on the SL or SPL to be measured at selected worker positions.

References
1. API RP 520. Sizing, selection and installation of pressure relieving devices in refineries.
Pt 1. Sizing and selection (March 1993)
Pt 2. Installation (December 1994)
2. Guidelines on noise. Medical Research Report EA 7301. API 1973
3. Design principles: Working environment. NORSOK Standard S-DP-002 Rev 1, Dec. 1994 (PO Box
547, N-4001 Stavanger, Norway. Fax (47) 51562105.)
4. 86/188/EEC Council Directive of 12 May 1986 on the protection of workers from the risks related to
the exposure to noise at work.
5. The noise at work regulations. SI No. 1790, 1989.
6. IEC 534-8-3: 1995 Industrial process control valves
Part 8 Noise Considerations
Section 3 Control valve aerodynamic noise prediction method
7. API RP 521. Guide for pressure relieving and depressuring systems (March 1997)
8. Noise Reduction. L L Beranek (Ed) McGraw-Hill (Pub. 1960)
9. Noise control for engineers in processing industries. Course notes 1990.
Institute of Sound and Vibration Research, University of Southampton, Hampshire, UK.
Appendix to safety valve noise; limits, reduction and control

A letter of enquiry
February 1996
Dear Sir,
LIMITS AND CONTROL OF NOISE FROM THE RELEASE OF SAFETY VALVES DIRECTLY TO
ATMOSPHERE
1. Philosophy for Safety Valve Noise
Currently, we have occasion to consider what philosophy we should adopt on future projects in regard
to the noise from safety valves which release directly to atmosphere, i.e. those that are not connected
to a flare or other system.
This letter, and your reply, will help define our future philosophy.
2. Limits to Safety Valve Noise
Limits to noise from "emergency vents" are often set at about 115 to 125 dB(A) at the ear of the
nearest personnel. See, for example, the 115 dB(A) of API EA 7301 (1973). Here 'emergency' relates
to foreseeable design situations such as safety valve operation and emergency depressurisation.
3. Request for Comment
We seek your comments on the possibility of purchase of,
say:1.
2.

a safety valve which does not exceed 115 dB(A), both at 1.5 m from the pipe vent and at
positions 1 m from the valve body and 1 m from the down-stream pipe;
a safety valve with associated silencer element, which does not exceed the limits given
above.

We also seek your comments on other possibilities for reduction or control of safety valve noise at the
nearest personnel. Here we have in mind by way of example,

a.
b.
c.
d.
e.
f.
g.

quieter valve designs;

quieter valve systems;
common silencers to a group of safety valves,
high level discharge, i.e. tall vent-pipes on individual valves or groups of valves, to take
advantage of increased distance and directivity;
injection of water to cool the process fluid, say air or steam;
water curtains to act as a barrier to sound;
permanently installed sound barriers near to the safety valves or the personnel positions.

The foregoing should in no circumstance be taken as recommended solutions by FWEL.

4. Responsibility for Calculation and Test
We seek your comments on the suggestion that a safety valve vendor should be responsible for both
the calculation to verify the noise from the safety valve and the test procedure and measurement of the
sound pressure level and sound power level. The same responsibility would remain with the vendor
where a safety valve and an associated silencer element were bought from the safety valve vendor.
At this juncture, we also request that you send to us notes and information on:1.
2.

your current calculation method and its justification;

the test basis of your current range of noise data for safety valves.

5. Noise at Platforms, Stairways, and Ladders

The difficulty and cost of noise control may be reduced if advantage is taken of the distance and
directivity between a safety valve vent and the nearest platform, stairway, or ladder (P,S,L). We invite
your comments on the proposal to provide vendors with details of both safety valve vent positions and
P,S,L positions in order that they may take appropriate advantage of the distance and directivity effects
of their proposed solutions.
6. General Nature of this Enquiry
Please treat this enquiry as a discussion document, which may be seen as " for the general good", and
thus need not be regarded as strictly confidential.
Yours faithfully,
MDGR.
Acknowledgements: This paper was published in the December 1998 issue of Valve World magazine.
Valve World wishes to acknowledge that 1) this paper is based on a paper first presented at the
Institute of Acoustics Conference, Windermere, Cumbria, UK, November 1997, and 2) that this paper
was first published in Hydrocarbon Engineering - June 1998
The API (American Petroleum Institute) is an important authority as regards standards and
recommended practices for the use of Safety Relief Valves in the petroleum and chemical industries.
For example:
1.
2.

API Recommended Practice 520 Part I - describes the design, sizing and selection of
components for pressure relief systems,
API Recommended Practice 520 Part II - contains guidelines for piping practices and methods
for determining the reactive force created during valve discharge,

3.
4.
5.
6.

API Recommended Practice 521 gives guidelines for evaluating causes of overpressure and
pressure relief systems,
API Standard 526 - provides an industry standard to manufacturers of flanged pressure relief
valves and includes a common set of installation dimensions,
API Recommended Practice 527 - provides a basis for testing and acceptance for set
pressure and seat tightness of pressure relief valves.
API Standard 2000 - describes venting atmospheric and low-pressure storage tanks.

To order these standards, visit API. Although it is not possible to view the standards via the website,
there are some very interesting 'Technical Interpretations' which can be accessed.
Useful standards can also be obtained from the American Society of Mechanical Engineers (ASME) at
www.asme.org. The following have been noted:

Safety valves for pressure vessels,

Specification standards for self-operated and power-operated safety-related valves,
Heating boiler safety valves accreditation package
Guide to European valves for control, isolation and safety,

In Europe, safety valves are covered by the Pressure Equipment Directive (PED - see also Valve
World box). Standards can be obtained via the National Members of the CEN, the European
Committee for Standardisation (www.cenorm.be). Of relevance to safety relief valves is the work being
performed by Technical Committee 69, Working Group 10, on Safety devices against overpressure:
PrEN ISO 4126- 1 Safety devices for the protection against excessive pressure - Part 1 : Safety
valves
PrEN ISO 4126- 2 Safety devices for the protection against excessive pressure - Part 2: Bursting disc
safety devices
PrEN ISO 4126- 3 Safety devices for the protection against excessive pressure - Part 3: Safety
valves and bursting disc safety devices in combination
PrEN ISO 4126- 4 Safety devices for the protection against excessive pressure - Part 4: Pilot
operated safety valves
PrEN ISO D 4126- Safety devices for the protection against excessive pressure - Part 5: Controlled
5
safety pressure relief systems (CSPRS)
PrEN ISO 4126- 6 Safety devices for the protection against excessive pressure - Part 6: Application,
selection and installation of bursting disc safety devices
PrEN ISO 4126- 7 Safety devices for the protection against excessive pressure - Part 7: Common
data
A further tool for locating standards is the NSSN: A National Resource for Global Standards
(www.nssn.org). The NSSN is described as: "A cooperative partnership between the American
National Standards Institute (ANSI) , U.S. private-sector standards organizations, government
agencies, and international standards organizations. NSSN's goal to become a leader in the provision
of technical data and information about important developments in a global standardization arena."